TOWARDS A HISTORY OF JAPANESE NARRATIVE MUSIC

著者 TOKITA Alison I.journal orpublication title

KYOTO CONFERENCE ON JAPANESE STUDIES 1994　III

volume .non01-03page range 233-244year 1996-03-25URL http://doi.org/10.15055/00003558

TOWARDS A HISTORY OF

JAPANESE NARRATIVE MUSIC

Alison I. TOKITA

Monash University

Introduction

Japan has maintained and elaborated a tradition of musical narration (katarimono) well

into the modern era. Extant styles such as koshiki, shomyo, heike, kowaka, gidayu, itchu, ozatsuma,

tokiwazu and kiyomoto offer to the researcher a variety of traditions spanning a long period of

historical development. In addition, there are many musical narratives in the area of surviving

folk culture. If we add styles which no longer survive, such as sekkyo-bushi, the range becomes

quite significant, though hardly able to compete with China, which still boasts over 350 styles

of musical narration today.

Although existing as separate musical and performance genres, these styles at the same

time form part of what can be called a continuous tradition of katarimono stretching back into

the mists of time. Narrative performance in early Japan was one way of preserving and

reliving history; it was "performed history". As a literary culture developed, some of these

genres came to be regarded and studied also as literature; others remained purely as

performing arts. Heike narrative is one genre which has been until recently considered as a

literary work with a single author. Similarly with works such as the Soga Monogatari and the "literary" genre of otogizoshi, and with sekkyo- bushi texts, all of which are now being

reassessed as the written versions of performance texts carried by a class of performers whose

role in the creation process was to interact with a tradition and its audience.

The comprehensive research of Japanese narrative music is still in its infancy. This paper

endeavours to establish an appropriate framework for a comprehensive study of Japanese

narrative music covering the major extant genres. Such research should be a synthesis of the

musical and textual aspects of the narrative. By concentrating on the musical analysis of those

styles which are still extant in performance it will be both a synchronic study, but at the same

time it will have a diachronic dimension through the historical treatment of the genres in

relation to each other.

Change within the narrative tradition as a whole took many forms over the centuries,

including the use of new instruments, new musical scales, indeed an overall elaboration of the

musical aspect of the narrative performance. There was also significant change in the

increasing use of written texts, and the development of musical notations. In addition, new

1-233

kinds of subject matter emerged, as well as the re- use of old themes with new twists.

Furthermore, new performance situations emerged, such as the use of narrative in the puppet

theatre and the kabuki theatre form the seventeenth century. From around this time too we see

the specialization of singer and instrumentalist, which testifies to an increasingly complex

narrative situation. Whereas in medieval narrative, there was one performer who had complete

control over his performance, in the Edo period it was a joint effort; straight narrative

changed to narrative with a visual element, and a concomitant loss of the ability for

improvisation in performance.

Despite all the changes just alluded to, there are certain structural principles which

continued to form the basis of all narrative music genres. This includes the use of formulaic

material. It is hypothesized that a major aspect of change is the changing meaning and

function of formulaic melodic material, or musical patterns, in the various stages of the

development of narrative music.

It is this aspect of continuity within the various changes in the narrative tradition which

forms the focus of this paper. Most examples will be drawn from two genres which have been

transmitted to the present day, heike narrative and gidayu narrative'. They can be taken to

represent respectively the narrative music of the medieval period and the early modern (Edo)

period.

Problems of research

Research has been carried out by many people on both the literary and the musical

aspects of most of the genres mentioned, but because cross- genre comparative studies of

narrative music are almost non- existent, many problems arise, not least of which is that of

terminology (this will be discussed in the final part of the paper). In addition, methodological

problems abound. These issues will be discussed first. At this general level, three problems

can be noted: the joint use of two kinds of data: historical data and contemporary

performance, and their respective limitations; the issue of orality versus literacy; and the

integration of the verbal and musical aspects of research.

The use of two kinds of data: written documents and extant performance

If you want to do musical study, there is not much choice but to concentrate on extant

performance. Some might object on the grounds that only historical documentation can show

us what musical performance was like in the past. I would reply that the study of the texts,

scores and other documentation is a different issue. However, it has to be admitted that each

of these kinds of data has its limitations. Documents usually only indicate the verbal aspect of

narrative performance. Notations when they exist are not sufficiently detailed or prescriptive

to be able to recreate the performed musical narrative. The limitation of extant performance

on the other hand lies in the fact that change occurs constantly, not only through the

1-234

emergence of new narrative genres in the tradition as a whole, but also within one genre over

time, as it is influenced by later genres. This means that mutual influences are far from clear

and that historical interaction is difficult to establish. A style like heike narrative which

originated in medieval period has certainly changed since then and been influenced by later

narrative and other musical styles. One cannot take it for granted that heike was like this

when joruri emerged from it around the fifteenth century.

Orality and literacy: the application of Parry-Lord theory

The study of narrative in performance was revolutionized by the development of the oral-

formulaic theory of Parry and Lord (Lord 1960; Foley 1988). The theory is concerned with

the concept of "structural slots", and a stock of verbal "slot fillers", at two structural levels:

formulas and themes. A formula is a group of words (forming a phrase or line of poetic text)

which is regularly employed under the same metrical conditions to express a given essential

idea. A theme is an idea expressed in several phrases or lines regularly used in telling a tale.

This leads to a highly formulaic style which enables the storyteller to compose the narrative

in performance in response to the particular performance situation.

The Parry-Lord, or oral-formulaic theory has already been successfully used in the

Japanese context by Yamamoto Kichizo (1988) on goze uta and Hyodo Hiromi (1990, 1992) on

zato biwa and its implications for heike narration. Yamamoto and Hyodo both demonstrated

that each performance is different, and can be said to be composed in performance using pre-

existing materials, in the form of verbal and melodic formulas, close to the Parry-Lord model.

The tradition of Japanese musical narration his its origins in oral narratives, such as

those which formed the basis for early writings like the Kojiki (712). In addition, examples of

oral narratives can still be seen in folk styles such as zato biwa and goze uta. These do not

have written texts (except some which have been recorded by folklorists but not used by the

performers), and are performed by blind (formerly itinerant and outcaste class) performers.

Genres such as heike and gidayu narrative have however had considerable interaction

with the upper class literate culture, and rely in varying degrees on written texts and notation

for their transmission and whose performance has become largely fixed, without

improvisation.

In spite of this, there remains in their verbal syntax and structure vestiges of an oral

heritage. The oral nature of Japanese musical narrative is suggested by their conspicuous use

of formulas, both in the text and in the melody of the sung and instrumental parts.

Orality of verbal and musical aspects of the narrative

If we conceive of the tradition of Japanese narrative music (katarimono) as one broad

type of performing art, we can discern three, perhaps four phases of development or types of

narrative, relating to different degrees of orality, associated with different eras of Japanese

1_235

culture: the era of purely oral or pre-literate culture, the era of chirographic culture, the era of

print media and the era of modernization.

The purely oral type or pre-literate phase is putatively the kind of narratives recited by

the kataribe professional narrators for the Yamato rulers before the importation or widespread

use of Chinese writing, but which formed the basis for the official chronicles of the Kojiki and

the Nihon Shoki. The oral narratives which are recorded in the Omorososhi of Okinawan culture

and the Yukar narratives of the Ainu culture are further examples of this type. Probably the

narratives of blind priest-like reciters in Kyushu (zato biwa) and the blind women balladeers

of Echigo (goze) can be included in this type, although they operated within a literate culture.

The chirographic culture of the medieval era led to interaction between oral narratives

and the literate elite culture, resulting in written texts and eventually also some kind of

musical notation. The heike narrative, for example, was involved in intensive interaction with

literate culture and the aesthetic not only of the warrior class, but also of court culture.

When the print media developed in the seventeenth century, the new literacy almost

supplanted the oral aspects of performed narratives, but "residual orality", to use Walter Ong

(1982)'s term, can still be found in both music and texts. In this stage, we can include gidayuu

and kiyomoto narrative.

Possibly a fourth phase, for the post-westernization era, should be added, when the

desperate need for preservaton of traditional forms led to their fossilization. This stage also

eventually benefitted from the development of secondary orality, with the emergence of

electronic aids to preservation and transmission in the form of sound recording and the

electronic mass media.

This paper is concerned with the changes between phase two and three, as represented

respectively by heike narrative and gidayuu. The orality of the verbal narrative started to be

influenced by the production of written manuscripts in the Kamakura period. There were "authorized versions" (called kataribon) of the performed heike narrative , such as that of

Kakuichi Kengyo in 1371. The Kakuichi version was intended to be an authorization of the

professional organization or guild (todo) newly established to protect the blind, and of which

Kakuichi was the head. It was a secret document, to be used for the transmission of the

orthodox line of the heike narrative. On the other hand, there were "reading texts" (yomihon ),

such as Genpei Seisuiki (Hyodo 1993), produced from the beginning to be read not recited.

When sighted amateurs began to learn heike recitation as a hobby from the sixteenth century,

texts with quite highly developed musical indications came to be produced. The most famous

and musically comprehensive one is the Heike Mabushi 1776. This became the standard for

performance in the Maeda school of heike narrative (heikyoku), which is the style of heike

narrative preserved to the present day in both the Nagoya and Sendai schools of recitation.

Joruri was originally an oral narrative art performed by itinerant outcast nun-like figures

(bikuni), as well as by blind male heike performers. When it became an urban commercial

entertainment in the form of puppet theatre with musical narration by sighted secular males,

its texts started to be printed. The earliest published texts (shohon or "certified editions") date

1-236

from the 1620s to the 1660s, and contain rudimentary musical markings (the names of musical

patterns such as kiri, sanju, and fushi ). By the 1770s there was a marked increase in the

number of pattern names. These shohon were for the use of professionals and were not

available publicly. When amateurs started to learn joruri from the 1680s, there was increased

demand for the publication of pattern names; one reciter, Harima no jo, permitted the sale of

texts showing musical patterns, which were circulated among amateurs (Tanaka 1984).

Thus, texts of joruri were used not only in the process of composition, but also in

performance, and in the process of transmission. The performers and students mostly wrote in

their own personal musical markings to help them in memorizing the music.

Writings about narrative music were also authoritative. For example, Uji Kaga no jo's

Takenoko shu (1678) gives in its preface and afterword a discussion of aesthetic aspects of

performance and instructions of various kinds for amateurs, including some explanation of

musical patterns. Another example is Takemoto Chikuma no jo (Takemoto Gidayu) in his

notes to a collections of pieces (danmonoshu) such as Joruri Toryu ko hyakuban (c. 1702-3).

Further, in Takemoto Gyoku Hiden (c. Hoei 1704-10), he gives some simple explanations of

named musical patterns, mentioning seven types of sanju, seven types of okuri, and harufushi,

fushiharu, iro, fushi, suete, toru, noru. Gidayu narrative also developed catalogues or lists of its

prominent musical patterns, called fushizukushi, from the 1780s. There were two types of

fushizukushi: one, a catalogue in the broader sense, which was a kind of practice piece; the

other kind were not for performance, and were indeed lists in the narrow sense. However,

these lists of named patterns do not include the majority of the commonest most basic patterns

of gidayu narrative, perhaps because everyone knew them anyway. According to Tanaka

(1984), unnamed patterns form the basis of gidayu narrative.

To what extent this use of written materials represents literacy versus orality is a matter

which still needs further examination. Even though texts and notations in varying degrees of

detail exist, the oral nature of narrative in composition, performance and transmission remains

very strong. This is particularly true of the orality of the non-verbal or musical aspect of the

narrative, which has persisted longer than the orality of the verbal (textual) narrative, since

musical notations have not played a major role in performance or transmission of Japanese

narrative music. More significant than written notations is the use of formulaic melodic

material, or musical patterns, as a means of preserving and transmitting the musical tradition,

and the naming of such patterns. It is possible that such material also played a role in the oral

composition of musical narratives in the purely oral stage.. It seems that named patterns

became a metalanguage for the music, whereby practitioners could talk about melodies and

manipulate them in the processes of teaching and memorizing, of composition, both in

performance in the oral stage, and in writing down the music alongside texts in the more

literate phase. Named musical patterns thus acquired some of the functions of notation, since

they were used alongside the texts to indicate what melody to use.

1-237

The integration of the verbal and musical aspects of research

Heike texts have to a large degree been worked into literary texts, and are not records of

a performed text in the same sense as when a folklorist collects oral narratives in the field.

Even so, there may be some justification for invoking the Parry- Lord theory to consider

vestigial oral elements. However, this theory was developed in the context of the verbal

aspect of oral narrative and does not attempt to account for the musical aspect. A theoretical

framework which will explain both the musical and verbal aspects of narrative performance

has still to be formulated. Of pressing concern is the question of the relation between

formulaic verbal expression and musical patterns which by analogy can be called musical

formulas. It is not clear whether they are any more than analogous, and whether the same

theory can be applied to musical formulas as to textual formulas. It is however suggestive

when considering the relationship between text and music. It is tempting to consider

stereotyped musical material as similar to textual formulaism.

Examples or a verbal formula in heike and other early narratives is the phrase saru hodo

ni, used frequently at the beginning of a new chapter (piece). An example of a theme from

heike narrative would be "dressing the hero", when there is a detailed description of the hero

(or other character)'s dress before an important scene. So it is necessary to consider whether

musical formulas can be found which correspond to such verbal formulaism.

An analogy can be drawn between verbal and musical formulaism in the following way.

The verbal formula at the level of the single phrase or line of text corresponds to phrase

length musical patterns, commonly called senritsukei. The longer verbal "themes" covering

several lines of text can be said to be analogous to sections in musical narrative, commonly

called shodan. Both of these musical units can be called "formulaic melodic material" in a

similar way to the verbal formulaism of the text, in that to a degree they can function as "structural slot fillers" .

Close analysis of this issue has still to be undertaken, but when one starts to look for

examples of correspondence between verbal and musical formulaism in heike and gidayu, it

begins to appear that the relationship between verbal and musical formulas is not necessarily

close. A couple of examples will be discussed. The style of narrative delivery which is closest

to speech, called shirakoe in heike and kotoba in gidayu, is not necessarily reserved for the

speech of characters in the dramatic sense. Particularly in heike, there is no relationship

between the speech of characters and the style of narrative delivery. In gidayu, on the other

hand, speech is commonly, but by no means always, delivered in the kotoba mode.

Another example of lack of correspondence between musical formula and the text is the

case of the heike formulaic melody sanju; musically this is something like a climax with its

high pitch register and melismatic delivery, but the text in sanju sections includes what can

only be seen as prosaic aspects of stating the time and place of the scene or action to be

described.

Historically, the elaboration and systematization of musical patterns seems to have taken

1-238

place late in the second phase (interaction with literate culture) or, more likely, early in the

third phase, when the printed media dominated , thus postdating the era of "composition in

performance" of the first phase of purely oral narrative, which forms the basis of the oral-

formulaic theory. For this reason too , therefore, the function of musical patterns can not be assumed to be the same as that of verbal formulas .

It is noticeable that verbal formulaism decreases from the medieval period to the early

modern (Edo) period. The texts of medieval narratives such as otogizoshi and sekkyo -bushi

are far more "old-fashioned" and formulaic , in fact "oral", than that of "modern" (toryu) joruri. Especially after Chikamatsu, the use of such phrases as saru hodo ni all but disappears from

joruri texts, except for the case of ozatsuma- bushi, which has a very archaic feel to it.

Formulaism did not disappear completely from joruri texts , but it was much less hackneyed. It happens therefore that at the very time when verbal formulaism weakened , musical

elaboration, in the form of elaborate musical patterns , occurred, then weakened again in the

phase of modernization. Or put in another way, orality gave way to musicality. As orality

declined, musicality arose.

The oral-forumulaic theory is probably of limited application to these musical narratives . However, the issue of orality itself, and the differential between verbal and musical orality , is of great interest. It is particularly relevant to the future and preservation of musical narrative

in Japan.

Framework for musical comparison

For the remainder of this paper , a framework for comparison of extant genres which originated in different periods will be undertaken , taking the example of heike as a representative of phase two, and of gidayuu as a representative of phase three . Many obvious differences can be pointed out between heike narrative , which originated in the thirteenth century, and gidayuu narrative which originated in the seventeenth century . These differences include a change in the accompanying instrument from biwa to shamisen, the social status of the performers, the predominant culture of the audience , the content of the narrative, and,

perhaps of most significance, the change in narrative situation from one of straight story

telling in private occasions and on street corners and in temples to one of dramatic

performance in the puppet theatre. In addition, on the musical side, there was a decreasing

reliance on formulas in both text and music, more complex musical scales and greater musical

elaboration.

While real and significant, these differences can mask underlying similarities of deeper

structure. We see important continuities, which allow us to speak of a "narrative tradition" as

an entity. These continuities lie principally in the basic structural principles of (mostly)

named musical patterns, at two levels: that of the section and of the phrase.

1-239

Structural similarities: vertical narrative structure

Broadly speaking, heike and gidayu narrative share what can be called vertical structure.

This has been called sekis-sei, or layered structure, by Yokomichi, who created this potent

model to explain the structure of one type of no play (mugen no) (Komada 1993).

Table 1: Vertical narrative structure

heike gidayuu

Long narrative Heike Monogatari Multi-act play

(jidaimono or sewamono)

piece "ku" (kyoku) dan ba

(daidan)

(dan) *section (dai) senritsukei shodan

shodan

kyokusetsu *phrase senritsukei senritukei

kyokusetsu kyokusetsu

motive: voice (senritsukei)

motive: instrument bachi/te to

(independent & linking)

The largest unit is that of the long narrative, which in the case of heike is also tha genre

itself. Joruri, originally the name of the story of the Princess Joruri, became the generic name

when the narrative content diversified.

The next level is that of independent units of narrative which can be performed

separately, and which correspond to the English term of a "piece of music" or kyoku in modern

Japanese. The traditional term for this level in heike is ku (literally "phrase"), and in gidayu,

dan, "section" or act of a play. In a sewamono (contemporary) play there are three maki instead

of the five dan of a jidaimono (historical play). Gidayu also has smaller dramatic units called ba

scene .

The next structural level is that of the section (shodan, literally "small section"). There

are apparently no traditional terms for this level, but Yokomichi's analysis has made the term

shodan widespread among researchers. Each section tends to have a unified character or style

of delivery (which can be called katarikuchi) and is closed off with a definite cadential

formula, making them discrete units.

At the next level is the structural unit of the phrase. Phrases are used in more or less

fixed sequences to form sections. Although a section may be characterized by a unified style

of delivery or other musical characteristics, not all of the phrases which constitute a discrete

1-240

section can be defined as musical patterns. This is why the concept of the narrative (sub)style

of the section as a whole is necessary. The conspicuous musical pattern is one characteristic

of this style, and the most conspicuous melodic patterns are cadential formulas.

These phrases in heike are for voice alone, since the biwa plays only in between vocal

phrases. In gidayu, some phrases are for voice alone, but most are for voice and shamise. In

heike, large independent instrumental solo phrases which precede sung sections, are called

bachi (sanju no bachi etc.), bachi meaning plectrum. In shamisen music, although also played

with a bachi (of somewhat different shape), such instrumental solo parts are called to (literally "hand"). Long introductory shamisen melodies are called maebiki (prelude), and those which

come between sections in the piece are called al no to (interludes). Because of their

independence, such instrumental preludes and interludes can sometimes be thought of as

sections rather than as phrases.

Smaller units than the phrase can be identified both in the voice and instrumental part,

but particularly in the instrumental part. Te can refer to a shamisen phrase or motive of any

length. The shortest ones which have some independent identity are several motives which

are used in between accompanied vocal phrases, such as tsunagi and kei.

Each of the layers or levels outined above tends to be formulaic in nature, but especially

important is the formulaic nature of the section and the phrase.

The conventional explanation of Japanese narrative music is that (working upward from

the bottom of Table 1) it is made up of previously existing phrase length musical patterns

which are cumulated until they form a section; the sections cumulate till they form a piece

(dan). This is often called a "mosaic", a term which implies a jigsaw like patterning, with a

random element; it is also non-linear, lacking a time element.

The vertical relations between the levels do not always seem to be exactly the same for

these two genres, especialy the case of the section and phrase. The table shows that the

terminology used by various researchers is not unified, even by researchers in one genre. This

presents problems for cross-genre comparison. Therefore, not only terminologically, but also

conceptually, the vertical relation between the levels is not easy to differentiate. For example,

in gidayu, sanjil is said to be a senritsukei (kyokusetsu), although a complex one; whereas in

heike, various researchers have called it different names: daisenritsukei (Kindaichi), kyokusetsu

(Komoda), kyokusetsukei (M. Gamo). Furthermore, Komoda adds levels of analysis in between

the levels of section and piece: dan (section) and daidan (large section). It is not entirely clear

whether sanju in heike is a senritsukei (kyokusetsu) or a shodan. It occupied in Nasu no

Yoichi almost eight minutes of singing (narration), being closed off with the cadential pattern

sagan (which always comes after the sanju). Therefore, this cadential pattern might be said to

be part of the substyle (katarikuchi) of sanju as a type of shodan instead of a senritsukei

(formulaic phrase).

Based on my preliminary analysis of these two genres, I would like to suggest this

working definition of shodan: a relatively independent unit of narrative, consisting of several

lines of text and several musical phrases, and which is closed off with a cadential formula. It

11-241

is therefore a self-contained section which tends to have a unified character in terms of

musical style, making it formulaic and predictable in character. This gives rise to a set of

different types of shodan in any given narrative genre.

My working definition of the Japanese senritsukei is a formulaic musical phrase

containing one phrase of text (usually 7 + 5 syllables in gidayu and other joruri), which has

little musical independece.

Further analysis is needed to see how far the concept of musical pattern is indeed the

same from one narrative genre to another. It may turn out that the relation between the

structural levels (i. e. phrase and section) is indeed different. Another area for cross-genre

comparison is varying degrees of pattern dependence. The proportion of patterned to non-

patterned material can be expected to be different from genre to genre.

Horizontal narrative structure

The discussion so far has focussed on vertical structure of musical narrative in a rather

static way. In addition, a linear dimension should be considered, the time axis of the

performance. This can be seen at each of the above levels, but let us look at the level of the

section and the phrase.

The cumulation of sections to form a piece is not altogether random. Just as in the

content of the narrative, there must be a beginning, a middle and an end, so too in the musical

structure, there are some types of section which can only be used at the opening of a piece, or

towards the beginning, and others which only can only be used towards the end. In addition,

there are certain types of section which are closely tied to a particular content, such as in

heike narrative the pattern hiroi for narrating martial scenes, and the pattern orikoe for

narrating lyrical romantic or pathetic scenes.

Within the section, with its unified style, there is still usually a development from the

beginning to the end in terms of the kind of phrases used. Most clearly, cadential patterns can

only be used to close the section, and to a lesser degree there are some special patterns which

are used for the opening of some sections. Some complex sections, especially in gidayu,

commonly begin in free rhythm in declamatory style and progress through a more rhythmic

style towards lyrical expressive aria-like singing at the end of the section, to finish with a

cadential formulae to give a sense of finality to section. This kind of linear progression, or

horizontal structure, is purely formal, not related closely to the content of the narrative.

Komoda (1993) successfully combines vertical and horizontal structure in her analysis of

heike narrative. It remains to be seen whether her method can be applied to gidayu and other

narrative styles.

Another area of musical continuity between heike and gidayu narrative is the progression

of the melody through different pitch areas, which serve as the basis of the formation of fixed

phrase length patterns. The heike patterns shoju, niju and sanju (first level, second level and

third level) originally referred in theory to three pitch areas, through which the melody

1-242

moved. Eventually, the melody associated with each area came to be formulaic , and became a fixed pattern. The same phenomenon can be seen in gidayu narrative , when pitch terms such

as naka, gin, haru, which originally meant to "start singing on (these) pitches", eventually took

on the character of musical patterns. This kind of similarity also indicates that the basic

musical structure and the approach to forming the narrative has been picked up or inherited

by the Edo period narratives from the medieval narratives , whatever other differences they may show.

Comparative analysis of heike and gidayu at the microlevel

The next stage is analysis at the micro-level. The problem of terminology and of concepts

has already been discussed. In addition, the following factors cause difficulty .

Heike as an older form is more systematized. As we have seen above , first its text, and later its music were systematized, in an effort to ensure consistency and orthodoxy . This

process in which orality moved towards literacy had an inverse relation to the popularity of

heike as an art, and also was witness to the separation of music from text . Gidayu, on the

other hand, is younger, and still visibly changing and developing. There has so far been very

little comprehensive musical analysis, despite the existence of catalogues of patterns

(fushizukushi) from the Edo period. Although patterns play an important part in gidayu, there

is much non-patterned material, and still room for improvisation.

Komoda's analysis of heike in terms of both vertical and horizontal structure is extremely

effective, but a comparable method of analysis for gidayu is still to be found.

Conclusion

The study to which this paper is a preliminary step - a musical analysis of contemporary

extant styles of musical narrative - will reveal more clearly the underlying continuities as

well as clarifying fundamental (as distinct from superficial) differences between them . It seems

appropriate therefore to focus on these two basic structural units for the comparative analysis

of narrative music: the key concept of musical patterns , and their combination in well-defined

sections (shodan) which have a unified style. It is hypothesized that these two units are the

key building blocks of all Japanese narrative music (katarimono) , though the boundary lines between them may be different in different genres. The focus on the formulaic section and

phrase as the major units of comparison will contribute towards the establishment of a theory

of musical patterns considering their role in composition , their role in transmission, and (if

any) their possible role in performance. In addition, light will be thrown on their relation to

the narrative text and their relation to textual formulas, in all these processes .

11-243

Notes

'Heike narrative is usually called heikyoku ("heike music") , the musical recitation of the Heike Monogatari (The Tale of the Heike). Gidayil narrative is also called gidayubushi, the musical recitation of

the puppet theatre (bunraku).The name comes from the name of Takemoto Gidayil who in 1684 set up his

own new style of joruri narrative in combination with the playwright Chikamatsu Monzaemon.

List of references

Foley, John Miles, The Theory of Oral Composition: History and Composition. Bloomington and Indianopolis:

Indiana University Press, 1988.

Hyodo, Hiromi, "Zato biwa no katarimono densho ni tsuite no kenkytt (I)" Saitama Daigaku Kiyo (Bulletin

of the Faculty of Letters, Saitama University) Vol. 26, 1990 (13-60).

Hyodo, Hiromi, "Zato (moso) biwa no katarimono densho ni tsuite no kenkytti (II)" Saitama Daigaku Kiyo

(Bulletin of the Faculty of Letters, Saitama University) Vol. 28, 1992 (35-78). Hyodo, Hiromi, "Yasaka-ryi3 no hassei: 'Heike Monogatari' to tekusuto ni okeru chttsei to kinsei". In

Ronshu: Chusei no Bungaku, edited by Kubota Atsushi. Tokyo: Meiji Shoin, 1994 (23-47).

Lord, Alfred B. The Singer of Tales. Harvard Studies in Comparative Literature, 24. Cambridge, Mass.:

Harvard University Press, 1960.

Komoda, Haruko, "Heikyoku no kyokusetsu to ongaku kozo" in Heike Biwa: Katari to Ongaku edited by Y.

Kamisango. Tokyo: Hitsuji Shobo, 1993 (161-193).

Ong, Walter, J. Orality and Literacy: the Technologizijg of the Word. London: Methuen, 1982.

Tanaka, Yumiko, Gidayu-bushi no senritsukei. Unpublished graduation thesis, Tokyo University of Fine

Arts, 1984.

Yamamoto, Kichizo, Kutsuwa no oto ga zazameite: katari no bungei ko. Heibonsha Sensho 122. Tokyo:

Heibonsha, 1988.

11-244